This Week’s AWS Outage And Why Blubrry Was Affected



If you use Blubrry hosting, you may have noticed a pause in service earlier this week as Amazon Web Servers went down worldwide, affecting hundreds of thousands of digital content providers – the vast majority of which were not Blubrry customers. You may also have been asking yourself, “What does Amazon hosting have to do with me – I’m not an Amazon hosting customer!”

If you wondered why an Amazon outage affected you, and how it happened in the first place, here’s the answer.

Like many web and digital services companies, Blubrry uses Amazon’s limitless storage system for our media hosting. At Blubrry, we use it because it provides endless storage and comes with a guarantee of service, which has been failure-free since we started using it in 2009. Tuesday, February 28 was the first time we have encountered an issue with the service that directly impacted you, our customer. Amazon has explained that the problem was due to an employee error and that the length of the outage was limited.

Our media delivery of fresh podcast content was not impacted by the issue, as we use a separate partner for our content delivery network. Statistics were not affected, nor were other key services we provide.

Here’s a list of what was impacted in the outage:

  • The ability to upload and publish new episodes was interrupted from 1:00 – 5:00 PM Eastern time.
  • There was limited access to archives of episodes between 1:00 and 3:00 PM Eastern time.
  • Customers experienced delays updating their podcast artwork in the Blubrry podcast directory.

We are currently reviewing the failure that happened at Amazon and will be incorporating changes to mitigate future failures so that podcasters are not impacted in the unlikely event that this problem occurs again. Amazon has also vowed to make changes to its services that will render such an event unlikely in the future.

Since AWS has such a long, solid track record of steady service without failure, we are Blubrry are confident that this was a temporary, if annoying, bump in the road.

Please follow and like us:

2 thoughts on “This Week’s AWS Outage And Why Blubrry Was Affected

  1. Your post contains some inaccuracies. “Amazon Web Servers went down worldwide” – Actually it was only in the Northern Virginia (us-east-1) data center. “Amazon’s limitless storage system” – it’s not really limitless, but virtually is. “which has been failure-free since we started using it in 2009” – there have been other minor service interruptions, though maybe not ones that you’ve noticed. Check here for the facts:

    https://aws.amazon.com/message/41926/
    https://status.aws.amazon.com/rss/s3-us-standard.rss

    1. Hello Ben,

      As far as the statement is concerned, we will stand behind it since as an outcome of this one service outage the perception world wide is that Amazon Web Servers went down. But again you are technically correct, it was technically only one service in one region that went down. It was not one data center though, it impacted the entire us-east-1 region, which is made up of multiple data centers. It did have a ripple effect, many other services at Amazon that use S3 were impacted. More critically, the AWS Service Health Dashboard, utilized world wide website to monitor the health status of AWS services was affected as well as the AWS console to access other S3 regions. Beyond that, many other services like ours were impacted world wide due to this one problem in one region of AWS data centers.

      You are correct that Amazon does not state anywhere that S3 storage is “limitless”, they do however say that there is “no limit” (http://docs.aws.amazon.com/AmazonS3/latest/dev/BucketRestrictions.html). I do not know what the significant difference is grammatically between “limitless” and “no limit” is in this context, but I am sure you are correct that the word usage may matter. I am confident though that Amazon has a plan in place to never allow a limit to be reached.

      Because we use a separate CDN partner with our services and our applications are designed to deal with expected issues with AWS services, we have never been impacted by past S3 issues.

Leave a Reply

*